COMPARATIVE ANALYSIS OF SPEECH INTELLIGIBILITY IN CHURCH

ACOUSTICS USING COMPUTER MODELING

By

SANGJUN LEE

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN ARCHITECTURAL STUDIES

UNIVERSITY OF FLORIDA

2003

I am dedicating my thesis to my parents.

ACKNOWLEDGMENTS

Many people have assisted me in the preparation of this thesis. My special

appreciation goes to my committee chairman as well as academic mentor, Professor Gary

W. Siebein, who has offered much encouragement and guidance throughout the course of

my research and the writing of this thesis. His knowledge and enthusiasm about acoustics

has certainly been an inspiration in my research.

I would like to thank the other members of my committee, Professor Martin A.

Gold and Bertram Y. Kinzey Jr., for their guidance and help. It was really a great honor

for me to be able to study with Professor Kinzey. His many years of experience in

teaching and consulting acoustics guided me to the right direction.

Thanks to these graduate students in the doctoral program, Hyeongseok Kim,

Bumjun Kim, Youngmin Kwon and Pattra Smitthakorn. Without their assistance using

acoustic computer program and their advice on writing my thesis, I could not have

finished this thesis at this time.

I would also like to thank to Hyun Paek and John Lorang who currently work as

acoustic consultants at the Siebein Associates, Inc. They helped me in many ways during

my practical training and finding data for computer modeling.

Most of all, my gratitude goes to my parents, my sisters and brothers-in-law. Their

support and understanding of my work offered me the opportunity to study at this level.

iii

TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................................................................................. iii

ABSTRACT....................................................................................................................... vi

CHAPTER

1 INTRODUCTION ........................................................................................................1

2 ACOUSTICS OF WORSHIP SPACES .......................................................................3

Historical Review of Church Acoustics .......................................................................3

Acoustical Design Goals for Worship Space................................................................5

Acoustical Design Guides for Worship Space..............................................................7

Room Volume .......................................................................................................7

Room Shape...........................................................................................................7

Noise Control.........................................................................................................8

Surface Treatments ................................................................................................9

3 ACOUSTICAL QUALITIES .....................................................................................10

Speech Qualities in Church Acoustics........................................................................10

Music Qualities in Church Acoustics .........................................................................11

Loudness (G) .......................................................................................................11

Reverberation ......................................................................................................13

Ensemble .............................................................................................................15

Warmth and Brilliance ........................................................................................15

Balance ................................................................................................................16

4 METHOD ...................................................................................................................17

5 RESULTS AND ANANLYSIS OF EXPERIMENT FOR CHURCH SPEECH.......21

Room Shape vs. Speech Intelligibility (Experiment 1) ..............................................21

Section Studies of Fan-Shaped Room (Experiment 2) ...............................................29

Section Studies of Rectangular Room (Experiment 3)...............................................34

iv

6 RESULTS AND ANALYSIS OF EXPERIMENT FOR CHURCH MUSIC............44

Loudness .....................................................................................................................44

Clarity .........................................................................................................................46

Reverberance ..............................................................................................................48

Intimacy ......................................................................................................................49

Spaciousness ...............................................................................................................50

Warmth and Brilliance................................................................................................52

Envelopment ...............................................................................................................52

7 CONCLUSIONS ........................................................................................................55

APPENDIX

A DATA FOR SEVERAL ACOUSTIC PARAMETERS IN DIFFERENT SHAPE ROOM MODELS.......................................................................................................59

B DATA FROM CATT ACOUSTICS-bASIC MODELS ............................................69

C DATA FROM CATT ACOUSTICS OF FAN-SHAPED ROOM .............................78

D DATA FROM CATT ACOUSTICS OF RECTANGULAR SHAPED ROOM WITH MODIFIED FRONT CEILING SHAPE....................................................................82

LIST OF REFERENCES...................................................................................................89

BIOGRAPHICAL SKETCH .............................................................................................92

v

Abstract of Thesis Presented to the Graduate School

of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science in Architectural Studies

COMPARATIVE ANALYSIS OF SPEECH INTELLIGIBILITY IN CHURCH ACOUSTICS USING COMPUTER MODELING

By

Sangjun Lee

August 2003

Chair: Gary W. Siebein Major Department: Architecture

The objective of this research is to investigate the relationship between room

features and speech intelligibility in church acoustics. Three different church room shapes

-- rectangular, fan, and round-- were modeled in AutoCAD and the models were then

exported to CATT acoustic analysis software. The comparison of RASTI values

estimated from CATT indicated that the fan-shaped room with a 60° inter angle had the

highest mean RASTI values, and the rectangular room had the second highest values.

However, the overall mean RASTI values were only between 0.42 and 0.45, which

indicated a ‘poor’ rating of speech intelligibility.

Additional acoustical treatments and modifications of room shape were performed

in the rectangular and fan-shaped rooms. When absorbent materials were placed on the

rear wall surfaces, the RASTI values increased to 0.57, giving a ‘fair’ speech

intelligibility. Ceiling shape modifications in the rectangular room resulted in mean

RASTI values over 0.60 at all receiver positions in the room, which indicated a ‘good’

vi

rating of speech intelligibility. Ray diagrams in each section model confirmed that strong

early sound reflections from the ceiling were a very important element for providing

optimum speech intelligibility. It was concluded that RASTI values are strongly related

to room surface configurations around sound sources and the acoustical characteristics of

the surface materials in the room.

In addition, musical sound qualities were investigated using several acoustical

parameters from the literature including reverberation, intimacy, spaciousness,

envelopment warmth, brilliance and clarity. According to the parameters for preferred

listening condition by Ando (1998), appropriate musical qualities of intimacy,

spaciousness, loudness, and reverberation were found in all of the rooms. However, the

rectangular room with a single concave ceiling had a clarity index (C80) value higher

than suggested in literature for music, but appropriate for speech clarity. It was found that

2.0 sec. of RT room should provide the simultaneous qualities of clarity for speech and

reverberance for the music.

vii

CHAPTER 1 INTRODUCTION

Acoustics is not a matter of quantity. Numbers from data don’t tell everything. Acoustics is a matter of quality-Bertram Kinzey, Jr. (2003)

One of the most important and ongoing issues on church acoustics is what are the

possible ways of designing an acoustical environment that is suitable for musical sounds

and speech. Acoustics for church music requires producing enough reverberant sound for

instruments and congregational singing, but speech sound requires only about one-half of

reverberation time for church music because it must be heard clearly to be understood.

Therefore, it is not easy to satisfy the acoustical requirements for these two environments

at the same time. Ideally one should not be sacrificed for the other (Berry & Kinzey,

1954).

One of the common resolutions for this issue in a contemporary church is

producing enough reverberation time for music sound and installing a sound

reinforcement system and hard surfaces near the speaker to improve speech intelligibility.

However, only highly sophisticated sound systems are appropriate, but they are not

necessary to use when a well designed architectural acoustic environment provides

natural sounds for ‘the realism’ and ‘the unity of the presence’ (Riedel, 1991). This study

was proposed to find possible ways to provide an optimum resolution of these two

different acoustic environments during the preliminary design process with a live

acoustic source.

1

2

The first step of this research compared the range of RASTI values measured by

CATT Acoustic analysis in different geometries of church rooms: rectangular, fan-

shaped, and round-shaped. Each computer modeled room had identical room volume,

congregational seating capacity, section, and reverberation time to evaluate only the

relationship between room geometry and speech intelligibility.

In order to obtain the desired range of Rapid Speech Transmission Index (RASTI)

values throughout the computer simulation, several different acoustic treatments on

surfaces and room geometry modifications were performed at each model. Ray diagrams

were used to find the relationship between the RASTI values obtained and the sound

energy distribution, which includes both the direct sound and reflections, at different

positions of the room. Ray diagram analysis is measured assuming a specular sound

reflecting; the angle of incidence of sound wave equals the angle of reflection where the

angle is perpendicular to the surface (Egan, 1988).

The second step of this research was to evaluate the music qualities of different

shaped church rooms. Loudness (G), Initial Time Delay Gap (ITDG), Inter Aural Cross

Correlation (IACC), Bass ratio (s), Treble ratio (s) and Lateral Fraction (LF) were

directly estimated from CATT Acoustics. These estimated values were compared in

different shaped room models and at 9 receiver different positions within the room. Each

parameter was compared with its preferred value in church acoustics or concert hall

acoustics.

CHAPTER 2 ACOUSTICS OF WORSHIP SPACES

‘Form follows function’ (Sullivan, 1896) is one of the most important philosophies

in the architectural design process. This means that a certain type of room must provide

for all of the functions that will take place in the facility, both now and in the future. In

worship spaces, preaching with the spoken word and liturgical music are generally

considered as primary parts of the act of worship. Therefore the acoustical environment

must be placed among the higher priority concern during the design process in order to

fulfill the main functions of a worship space (Riedel, 1983).

However, different religions have very different acts of worship. Unlike a common

western church service, prayer and meditation are the main functions in a temple in

Korea. When people need to sing or pray together on a special day, they get together in

the front yard of the main building rather than sitting inside the building. It is apparent

that the main function of a temple in Korea is not always singing and preaching.

The acoustical requirements of the room are determined by what will take place

programmatically within the space. However, these requirements have changed as time

has changed not only architecturally, but also liturgically. Therefore, it is necessary to

understand the changes of the acoustical environment in church design.

Historical Review of Church Acoustics

In the middle ages through the Renaissance, large Roman Catholic cathedrals had

highly reverberant acoustic environments for liturgical music and chants. However, the

3

4

early synagogues, which emphasized sermons and scriptural readings, required less

reverberant spaces than the medieval cathedrals.

The Protestant Reformation was begun in the beginning of 1517. It was a

movement which broke up the institutional unity of the church in Western Europe and

established the third great branch of Christianity. Since the Reformation time period, high

speech intelligibility has become more common and important requirement for the church

service than in the medieval cathedral (Lubman, 1983).

In the early 1940s, some churches were built in isolated places, and it was not

necessary to worry about acoustic problems such as noises from outside activities,

mechanical installations or even interior human activities (Buglio, 1992). After World

War II, many countries such as Korea and Japan, which experienced terrible war, needed

to rebuild worship spaces and started to use sound reinforcement systems (Nagata, 2001).

At first, this application seemed to be financially successful because inexpensive sound

reinforcement systems were used. However, people started to realize that they did not

have adequate acoustical equipment (Buglio, 1992).

As increased number of service participants, much bigger organ facility and bigger

room size were required. Therefore, the architectural aspects of church design also started

to change. At the same time, many types of mechanical equipment and new musical

instruments, such as electronic organs and synthesizers, were installed while traditional

instruments like the pipe organ and piano were still in use. These made it possible to start

to invest in acoustical quality of churches and the established need for acoustic experts to

assist in achieving not only proper architectural designs for natural acoustics but also the

proper installation of sound reinforcement systems (Jarzabkowski, 2002).

5

Modern worship spaces involve many different types of functions based on their

liturgical style. The highly reverent acoustics of a cathedral are well suited to the organ

and Gregorian chants. However, the simultaneous need for high speech intelligibility is

also important for cathedral style of church. On the other hand, an evangelical style of

church has as large a volume as a cathedral but contemporary popular music is usually

played with an electronic organ (Lubman, 1983).

Contemporary church auditoriums require various functions. A church space is no

longer only for worship but also for many other activities such as classes, meetings, and

recreation. These different functions need different acoustical conditions. Each room

should provide optimum acoustic parameters for the functions it will have.

Acoustical Design Goals for Worship Space

Modern worship services require very active interaction between the preacher, the

performing musical group and the congregation. All of the participants should be

considered not only for their own functions but also in dialogue with each other. The

main goal of worship service is to provide God’s message to all participants in the

service. The preacher primarily leads the service with speech. Thus, it is necessary to

produce loud enough speech sound so that it can be heard at any seat in the building. If

the speech is not heard properly, the participants may feel a sense of alienation. Everyone

in the church service must be in an equal acoustical environment because everyone is

equal in the house of God (Riedel, 1983).

The type of music and singing occurring in each church depend upon the liturgy,

theology and worship style of the religions. Therefore, the location of choir, choral or

praise band and its acoustical treatment must be considered in different ways. The general

consideration for the music performance group is that the all sound they produce must

6

travel through entire worship space. It is also important for the choir and band member to

hear sounds they produce, otherwise the members feel like they are singing alone

(Acoustic Sciences Corporation, 2002).

The listeners in the worship space have a dual function. As illustrated at figure 2-1,

they are not only sound receivers but also sound sources. At first as sound receivers, it is

essential for them to clearly receive speech sound from the preacher and singing from the

music performance group. This helps the congregation feel more engaged in the service.

As sound sources, the congregational singing is the most important church music, and the

most important function of the instruments such as organ and band is to lead and

encourage the congregation in singing, because congregational singing is the way they

respond to the God’s message (Berry & Kinzey, 1954).

The principle of acoustical design to provide the communication among preacher,

music performer, congregation and instrumental accompaniment is that all of these inter-

act each other simultaneously.

Organ

Choir

Preacher Congregationas sound receivers

Organ

Choir

Congregationas sound sourcesPreacher

Console

Console

Figure 2-1: Section drawing for the interaction between sound sources in church rooms.

Congregations can be both sound sources and receivers

7

Acoustical Design Guides for Worship Space

Room Volume

The room volume is one of the important elements that determine reverberation

time. This must be decided depending upon the type of religion and the number of

participants in the room (Jarzabkowski, 2000). For example, as shown figure 3-1, Roman

Catholic churches require a much longer reverberation time than protestant churches.

Therefore, Roman Catholic churches need a larger volume than protestant churches. In

addition, in church acoustics, the congregation is the major sound absorbing element. For

this reason, as the number of participants increase, the room volume should be increased

relatively to maintain desirable reverberation.

Room Shape

In church design, fundamental room shape and volume must be primary design

considerations to provide an optimum acoustic environment. Some shapes are more

beneficial than others to project early speech sound energy which is important for

optimum speech intelligibility. Some of them, however, may cause serious acoustical

problems such as standing waves, flutter echo, sound focusing and intensive late

reflections (greater than 100 ms). Concave shapes that concentrate reflected sound should

be avoided if it is not well treated in terms of sound diffusing materials (Riedel, 1986).

Long and flat parallel walls cause undesirable flutter echo, which decrease speech

intelligibility seriously. Spatial separation by alcoves, archways, and objects, which are

seen in many older churches, will diminish the effectiveness of early arriving sound

energy and weaken the speech intelligibility. This geometrical consideration must be

given not only in design of the building floor plan but also in design of the section as

well. Therefore, the ratios of the room’s length to its width as well as the ratios of room’s

8

height to width are the important considerations that must be made (Ingram, Barnes,

Hampton, Klepper, & Noehren, n. d.).

Figure 2-2: Section drawings of sound focusing from concaved-shaped ceiling and

convex forms within an overall concave area to avoid sound focusing Sources: Architectural Acoustics, Egan, 1988, p116 (Left) Acoustics in the Worship Spaces, Riedel, 1986, p22 (Right)

Noise Control

Modern worship services are confronted with many types of noise problems. On a

busy Sunday, car noise, people talking, and even airplane noise can intrude into the

church room through a single layer of window, wall, door or roof. Noise from mechanical

equipment, can cause serious reduction in the intelligibility of pastor’s speech. Therefore,

the human voice is no longer able to compete with these ambient noises without sound

isolation treatment (Doelle, 1972). Noisy equipment should not be placed next to the

church auditorium. Acoustical design of the HVAC system design is one of the possible

ways to reduce these ventilation noises. In addition, the material of auditorium shell must

be chosen carefully and constructed with highly sound isolating materials for the

congregation in order not to be disturbed by intruding outside noises (Siebein, 2002).

9

Surface Treatments

Sound travel leaves the sources in many ways. It is reflected, diffused or absorbed

by the various surfaces it encounters. All of the room surfaces need to be treated in

relation to the characteristics of the sound sources and receivers. The church room

usually requires highly reflective surface materials around the altar, organ, choir, band,

and congregation area. These reflections are necessary not only for the distribution of

both speech and music from the preacher, choir, and band but also to give a sense of

shared support to the congregation.

In addition, sound-absorbing treatment can also be placed in order to control

excessive reverberant sound energy or prevent acoustical problems. However, this

treatment must be located carefully in order not to absorb useful early sound energy from

sound sources. Therefore, highly sound reflective materials are desirable on the side walls

and ceiling. Usually, the rear wall is the most desirable place for sound absorbing

treatment (Ingram et al. n. d.).

CHAPTER 3 ACOUSTICAL QUALITIES

It has been possible to control the reverberation in rooms since Sabine’s formula

allows one to estimate a reverberation time of rooms. However, reverberation time is not

satisfactory to evaluate other acoustical qualities of the room. In the past several decades,

much research has been dedicated to evaluating the acoustic qualities of rooms in both

objective and subjective ways (Soulodre & Bradley, 1995). It has been considered that an

optimum reverberation time and speech intelligibility in various types of church

auditoriums is the most important acoustical parameters. However, unlike concert hall

design, the other acoustical parameters such as envelopment, intimacy, or spaciousness

are not generally dealt with in church acoustics.

Speech Qualities in Church Acoustics

Speech intelligibility in rooms is influenced by both the level of speech and

ambient noises and acoustic characteristics of the room. Higher speech to noise ratios

usually result in greater intelligibility of speech (Reich & Bradley, 1998).

The Hass Effect explains that the speech sound in an early arriving reflection is not

heard separately from the direct sounds because our hearing system integrates these early

reflections and direct sound together (Hass, 1972). Lochner and Burger (1964) also

found that early energy arriving within 50ms of the direct sound is more affective in

improving speech intelligibility than later arriving sounds. Therefore, speech

intelligibility is more related to the early decay time (EDT) than to the reverberation time.

The speech transmission index (STI), the rapid speech transmission index (RASTI),

10

11

which is simplification of STI, and the useful to detrimental sound ratio (U80) are

available to evaluate the speech intelligibility of room. STI is derived from modulations

function that considers both speech to noise ratio and the acoustic characteristics of room

(Bradley, 1988). ‘U 80 is the ratio of the direct sound and early reflections (up to 80ms

after the direct sound) divided by the reverberant energy plus the background noise level’

(Siebein, Crandell, & Gold, 1997, p40).

Music Qualities in Church Acoustics

There are several approaches to evaluate musical quality in a room. Each approach

needs to analyze one or more acoustical parameters for its objective evaluation. In church

acoustics, however, only a few of these parameters such as loudness, or reverberation are

usually discussed compared to the musical quality of the concert hall or any other type of

music hall is discussed. This chapter mainly focuses on acoustic parameters to evaluate

and optimize music quality in worship spaces.

Loudness (G)

Loudness is the subjective response of the audience to sound pressure or sound

intensity and it is the primary acoustical concern in most auditorium and concert hall

design. Loudness (G) is defined as the difference in sound pressure level at the receiver

location and the level of the same source an anechoic chamber 10m (33ft) from the sound

source in the room (Soulodre & Bradley, 1995).

Gp t dt

p t dtA

=

∞

∞∫∫

102

02

0

log( )

( ) [ : SPL measured in anechoic chamber] p tA ( )

In a church auditorium, loudness is important to achieve adequate communication

between the sound sources - speech (pulpit, altar, ambo, and speaker) and music (organ,

choir singers, and instruments) - and the sound receiver (congregation).

12

The first way to provide proper loudness is to raise the sound source level.

Therefore, the minister or priest in the pulpit and choir must emanate adequate sound to

every part of the auditorium without using the sound reinforcement system, because the

distance increases between these areas, air molecules and any type of sound absorbing

element attenuate sound energy (Knudsen, 1950).

The second way to improve loudness is to surround the chancel and pulpit with

reflective walls and ceiling in order to push the sound towards the congregation. This can

reinforce the direct sound and reflected sound together (Egan, 1988). According to the

Hass Effect, this reinforced sound can enhance loudness. Over -hanging canopies at the

chancel and pulpit area can reflect sound to the rear of the room where the sound is

remarkably weaker than front area; this is following the inverse square law of physics.

Out door:

I Wd

=4 2π

[I = sound intensity (W/m2), W = sound power (W), d = distance from

sound source (m)]

Reference: Architectural Acoustics, Egan, 1988, p10

Indoor:

Lp Lw QD R

= + + +104

4 102log[ ]π

[Lp= the sound pressure level (dB), R= room constant in Sabines of absorption

( sα∑ ), D= distance from the sound source (ft), Lw = sound power

level of a source (dB), Q= directivity]

Reference: Siebein, et al., 1997, p36

Well-elevated chancels and pulpit areas enhance not only the sightlines but also

prevent direct sound absorption by the congregation (Doelle, 1972). When sound is

13

absorbed by the congregation, the overall sound level is decreased. In large rooms, like

churches, the sound should move in the room to optimize the acoustical environment for

both the natural room acoustics and the audio system (Siebein, 2002). Of course, a sloped

seating is another approach to have proper sound propagation through out the church

auditorium.

Reverberation

Reverberation is smooth sound energy decrease after the successive reflections, but

they not perceived individually. The physical expression of this parameter is called RT60,

which is measured from the time it took the sound to decay 60dB of its initial sound

level. Sabine formula provides the quantitative relationship between reverberation time,

the volume of the room, and the total amount of the absorption in the room enclosure

(Doelle, 1972). Since W. Sabine defined his formula, it is possible to calculate

reverberation time with relative accuracy (Siebein et al., 1997). However, the Sabine

equation only applies to rooms in which the sound is diffuse so it should be used with

caution. When the sound is not diffuse, the sound decay curve is not ‘straight’ and the

difference between the observed and calculated reverberation time could be noticeable

(Knudsen & Harris, 1950). In addition, Egan said when the ratio of absorption to the

room volume is very high, the Eyring formula must be used.

The optimum RT60 for an auditorium is related to both the room volume and the

amount of absorption in the room. In a church auditorium, reverberation time over 2.0

(Egan, 1988, p133) is required for organ music and general church music, which should

have a longer RT than an equivalent volume room used mainly for speech. However,

participants use church auditorium for a number of different functions such as speech,

14

music, drama, and meeting. Therefore, the most important element must be decided, and

it should be a priority concern.

Sabine Formula:

T = 0.05 vA

[v= room volume (ft3), α = absorption coefficient, A S= ∑ α ]

Eyring Formula:

T . vS . ( a

=− × −

0 052 3 1log )

α α α α α= ∑

∑=

+ + ++ + +

SS

S S SS S S

n n

n

1 1 2 2

1 2

L

L = Mean sound absorption coefficient

[v= room volume (ft3), S= total surface area (ft2)]

Reference: Architectural Acoustics, Egan, 1988, Appendix A, p391 Table 3-1: Recommended RT (s) in different church religions vs. room size. Reference:

Environmental Acoustics, Doelle, p56 (NA: Not Available) Roman

Catholic Protestants Evangelical

church

Optimum RT mid sec Small (20,000 cu ft) 1.7 1.2 NA Medium (50,000 cu ft) 1.9 1.3 NA Large (100,000 cu and up) 2.0-2.6 1.4-1.6 NA

Small volume churches usually have seating capacities less than 400 or 500 people.

Such a small room has a short distance between sound source (pulpit, choir, and organ)

and sound receiver (congregation), and it can achieve intimate sound. If the room volume

is too small to have a long reverberation time, which is required for cathedral acoustics,

room surfaces must be treated with highly reflective material, and a very minimum

amount of sound absorbing materials should be used (Ingram et al. n. d.).

The medium size of church, seating capacity between 500 and 900, has a large ratio

of volume to seating area. It is large enough to have the space for choir, organ, and

organist. Large churches have usually a large volume and enough reverberation time for

15

music and singing, but excessive reverberation time can reduce speech intelligibility.

Therefore, sound absorbent material can be installed to provide optimum reverberation

time based on its volume. A sound reinforcement system is possibly installed over the

congregation to overcome speech intelligibility due to the excessive reverberation time,

and it also can be installed to overcome lack of live speech level to congregation more

remote from pulpit (Ingram et al. n. d.).

As stated chapter 2, the optimum RT60 may be also classified with its usage such

as catholic, protestant and modern evangelical churches. Performing romantic music

should have longer reverberation periods than rooms primarily designed for performance

of baroque or classical literature (Lubman, 1983).

Ensemble

Choir and band members in most churches are standing upright in straight wide

rows, one behind the other. It is unnatural and may be a problem because the sound from

choir and band members can be absorbed by the people in the front line. Therefore

singers must be located on risers so that their voices are not absorbed by the front row

singers (Riedel, 1983). Kinzey and Berry (1954) suggested that choir members must face

the console director, and be placed in the same acoustic environment as the organ.

Reflective walls that are behind and beside musicians and highly diffused surfaces

may scatter sounds and produce a blend of total sound from choir and band members.

This make possible for members to hear each other and even strengthen their sounds to

the congregation (Acoustic Sciences Corporation, 2002).

Warmth and Brilliance

Since an organ produces one of the lowest frequencies of the sound than any other

instruments, it is important to maintain its original sound through the church auditorium.

16

To reflect low frequencies of sound from an organ, sidewalls area, which is located

around an organ, should be approximately four times larger than a wavelength of the

lowest pitch of an organ pipe (Egan, 1988).

Hard and heavy materials are appropriate to prevent absorption of the low

frequencies of sound. The traditional medieval churches made of large and heavy stones

normally produce warm sound from the organ. These days, hard plaster, multi-layered

gypsum board and concrete are widely used (Ingram et al. n. d.).

Balance

Material balance: the ceiling is usually the main distributor of both music and

speech sound through out the church. Heavy and thick materials such as hard plaster,

several layers of laminated dry wall, or concrete are preferred to use at ceiling area. If

hard but thin materials are used, it may possibly absorb low frequencies of sound, making

a room sound unbalanced.

Sound balance between choir and congregational singing: the scattered sound

around choir area may enhance the balanced singing among choir members. A sense of

singing together among congregation members can be achieved by controlling sound

diffusing panels over the congregation ceiling area or sidewalls around them (Ingram et

al. n. d.).

CHAPTER 4 METHOD

The primary purpose of this research was to investigate the relationship between

speech intelligibility and the architectural features in the church room. Three different

church rooms: rectangular, fan shape, and round shape were derived from a review of

multiple church projects reviewed in a private acoustic consulting firm. The architectural

features of these three churches were simplified for the computational analysis. The room

volume and seating capacity were also modified in order to provide the same conditions

in each model room with the different shapes, because when rooms have the same room

volume and seating capacity, they have similar RT to each other.

The room volume and seating capacity were determined so there will be enough

sound pressure level throughout the room without sound reinforcement systems.

Generally the room capacity less than 1000 seats is acceptable for churches which

provide only natural sound sources or both natural sound sources and sound

reinforcement systems (M. David Egan, 1988). Therefore, seating capacity about 670

seats and room volume 5900m3 were determined in this study.

The three churches had identical section shapes with flat ceiling and surface

materials so that could be investigated the relationship between the room shapes and their

RASTI values. The following materials were considered for appropriate materials of

church room acoustics and used in this research.

• Floor: concrete floor • Hall way: acoustic tiles • Front and side walls: plaster on brick

17

18

• Ceiling: hard plaster • Rear wall: concrete block • Congregation: Fully occupied congregations seated in wooden pews

AutoCAD v2000 was used to design all the computer models and input all the

surfaces materials. Since all of the room models were symmetrical, half of the rooms

were made and exported into CATT Acoustic analysis v8.0. The interface functions using

Auto-LISP in AutoCAD are very helpful and efficient way to make 3D models and

export files into CATT Acoustic.

The sound source data, receiver locations, and surface materials can be modified

either in AutoCAD or CATT Acoustic. In this research, male loud speech sound pressure

level was used as a natural human voice. To calculate the RASTI values with background

condition, NC 25 dB was used to provide desirable speech to noise ratio in church

acoustics (D. Egan, 1988). The sound source was placed 2m set back from the edge of

altar area, and raised 1.8m from the altar floor. The directivity of this sound source was

derived from Egan, 1988.

The 9 receiver locations were decided to uniformly occupy half of the church room.

Each room was composed of three position rows: front, middle, and rear, and each row

was divided by three parts: right, center, and left side.

Table 4-1: Sound source and back ground noise level (dB) used in the CATT Acoustics analysis. Octave Band Centre Frequencies Hz 125 250 500 1000 2000 4000

Male Loud Speech SPL dB 50 60 68 65 59 56 Ambient Noise (NC25) dB 44 37 31 27 24 22

19

Figure 4-1: Speech Contours (Architectural Acoustics, Egan, p 83)

Figure 4-2: Sound source and receivers positions at rectangular room with single concave

ceiling

The second part of this research was to investigate the music quality of church

rooms. Omni - directional sound sources were used, and several acoustic parameters were

directly derived from CATT Acoustics. In this study, some of parameters (G, C80, EDT,

RT, and LEF) were directly measured from CATT Acoustic analysis, and some of them

(Bass ratio, Treble ratio, ITDG, and IACC) were calculated based on RT values at each

representative frequency or plotted graphs from CATT. Optimum values for these

parameters in a church environment haven’t been generally discussed. Therefore

20

recommended values for each acoustic criterion were compared with general concert hall

music recommendations.

‘Auralization’ through CATT Acoustic was performed to obtain the Interaural

Cross Correlation (IACC) values. ‘Auralization’ is the process where predicted octave

band echograms are converted to binaural impulse responses (D’Antonio, 1988). In this

study acoustic guitar sound from CATT anechoic sampling was used as a sound source.

CHAPTER 5 RESULTS AND ANANLYSIS OF EXPERIMENT FOR CHURCH SPEECH

The following abbreviations are used for each room

1. REC: Rectangular shape with flat ceiling 2. FS: Fan shaped room inter angle (90°) with flat ceiling 3. NF: Narrow fan shaped room inter angle (60°) with flat ceiling 4. RND: Round shape with flat ceiling 5. SCC: Single concave ceiling at front room 6. SCV: Single convex ceiling at front room 7. DCC: Double concave ceiling at front room 8. DCV: Double convex ceiling at front room 9. TCC: Triple concave ceiling at front room 10. TCV: Triple convex ceiling at front room 11. RDV: Double convex shape diffusive between rear ceiling and upper rear wall

Room Shape vs. Speech Intelligibility (Experiment 1)

Three different shapes of churches were used to evaluate the relationship between

room shape and speech intelligibility. RASTI and STI values were estimated from CATT

Acoustic, V8.0.

Table 5-1 shows a description of each computer room model. The materials, sound

source, and background noise level used in the computer simulations were discussed in

chapter 4. Each room has a very similar RT in the mid frequencies of approximately

1.95s. The congregational seating capacity of approximately 660 was used in each model

in order to provide the same absorption by congregation. Generally, spaces seating

between 500 and 1000 people may need sound reinforcement systems depending on their

spatial purpose. Spaces that seat more than 1000 people usually require sound

reinforcement systems (Egan, 1988). Therefore, using a room with less than 1000 seats

makes it possible to produce enough speech sound pressure level without sound

21

22

reinforcement system in church acoustics. The desired RT in a Catholic Church is

between 1.8s and 2.0s when it is fully occupied; the room volume is approximately

6000m3 (Doelle, 1972).

Table 5-1: Descriptions of three church room models (RT mid means that the average of reverberation time at 500Hz and 1000Hz)

Plotted by CATT Acoustic V,8.0

REC FS RND Volume(m3) 5893 5844 5908 Seats 672 656 652 Volume/seats 8.8 8.9 9 RT sec. mid 2.03 1.92 1.98

Measured RASTI values in three models were compared in figure 5-1. The mean

values in each model range from 0.42 to 0.45 where the subjective quality is judged

‘poor.’ The fan shaped model had the highest mean value of 0.45 among three room

models. However, the mean values at different positions varied in fan and round-shaped

rooms.

The standard deviation of RASTI values in the fan-shaped and the round-shaped

room was 0.02 and 0.04 respectively, but it was only 0.01 in the rectangular room. When

RASTI values were compared at each position, the highest mean RASTI value of 0.47

was found at the front positions of the fan-shaped room. At these positions, the

rectangular and round-shaped rooms had the mean RASTI values of 0.42, 0.43

respectively.

23

RASTI VS Room shape

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Front Middle RearReceivers positions

RAS

TI

0.00

0.15

0.30

0.45

0.60

0.75

0.90

excel

good

fair

poor

bad

RectangularFan shape(90')Round shapeFan shape (60')

Figure 5-1: Comparison of RASTI values at 9 receiver positions in different shapes of

churches

The inter angle of the fan-shaped room was changed from 90° to 60° , and RASTI

values were estimated in the narrow fan-shaped room. It was clear to see that the narrow

angle (60°) of fan-shaped room had the highest mean RASTI values of 0.50 at the rear of

the room, which indicates a ‘fair’ subjective judgment, but the middle and the rear of the

room had still a ‘poor’ subjective perception of speech intelligibility with the mean

RASTI values of 0.46 and 0.42 respectively.

Figure 5-2 shows that the RASTI variations in the rectangular and fan-shaped

rooms (60°) could be increased with different acoustical treatment such as sound diffusive

panels on the side walls, sound absorbent material on rear walls, canopy above pulpit,

and composite design of all three treatments (Figure 5-4).

24

Acoustical treatments VS RASTI (Rectangular model)

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Front Middle Rear

Receivers positions

RAS

TI

0.00

0.15

0.30

0.45

0.60

0.75

0.90

excel

good

fair

poor

bad

Witout treatment ABS at rear

Diff at sides Canopy above pulpit

Compositive

Acoustical treatment VS RASTI (Fan shaped room 60')

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

Witout treatment ABS at rear

Diff at sides Canopy above pulpit

Compositive

Figure 5-2: Comparisons of RASTI values in rectangular and fan-shape room with

several different acoustical treatments.

The standard deviation of RASTI values at different positions at both rooms were

0.035 and 0.03. The additional treatment on the room surfaces or ceiling resulted in the

increased RASTI values in both models.

25

The greater change in mean RASTI values was found in the room with composite

acoustic treatments and the second highest was found in the room with a sound absorbent

material on the rear wall. In the rectangular room, the mean RASTI values changed from

0.44 to 0.55 with composite acoustic treatments. The mean RASTI value was 0.51 with

sound absorbent material on the rear wall of the room. This was explained by the fact that

the sound absorbent materials on the rear wall could prevent later arriving sound energy

from the rear wall and flutter echoes that could occur between the front and rear walls.

The second most efficient way to improve speech intelligibility was found in the

room with the canopy above the pulpit area. The mean RASTI value only increased 0.03

compared to the mean RASTI value increase of 0.07 with sound absorbent material added

to the rear wall. The canopy above the sound source can reflect some of the early sound

energy, but the amount of reflected early sound energy is not sufficient to create ‘good’

speech perception. This is explained by the fact that there are higher Early Energy

Fraction values (D-50) in each octave band frequency in the room with the canopy than

in the other room. D-50 is defined as the early energy level, which is up to 50ms after

direct sound, divided by the total sound energy in the room (Siebein et al., 1997).

Therefore, the more D-50 mean value the more early energy exists in the room.

In this study, as shown figure 5-3, when the canopy were suspended over the pulpit,

D-50 mean value at 1K increased overall only 6.1%, but when the front ceiling shape was

changed to a single concave ceiling shape in the rectangular room, D-50 values increased

by approximately 19% and 29% at the front and the rear positions of the room. At the

middle positions of this model, the mean D-50 value increased to 39%. The mean D-50

26

value at IK with canopy increased from 46.3% to 52.4%, but it increased to 75.2% with

the single concave ceiling shape at the front of the room.

In addition, different directional distribution of speech sounds in the vertical plane

could explain this result. According to Kuttruff (1991), sound pressure levels of

frequency band at 1400 to 2000 Hz decreased approximately 5 to 7 dB as speech sounds

go up from the frontal direction 0° to the vertical direction 90°. Therefore, the canopy

above the pulpit only reflect relatively low intensity of speech sounds and this resulted in

the increase RASTI values only 0.3. However, the efficiency of the canopy can not be

judged in this study, because the acoustical effects of canopy may depend on many other

factors such as size of the canopy, suspended locations, and surface configurations.

It was concluded that the model with composite acoustical treatments had both

acoustical advantages of the suspended canopy above the pulpit and the sound absorbent

material on the rear wall. This is why the highest RASTI values were found in the room

with composite acoustical treatments.

D50(%) comaprisons at rectangular rooms

0102030405060708090

100

Front Middle RearReceivers positions

D50

(%)

RECREC + CanopySCC

Figure 5- 3: Comparisons of D50 (1K) values at three different rectangular rooms: flat

ceiling (REC), flat ceiling with canopy (REC +Canopy), and single concave ceiling (SCC)

27

The least efficient way to improve speech intelligibility was added sound diffusers

on the side walls. When sound diffusers were added on the side walls, figure 5-2 shows

that the mean RASTI values increased only 0.005 and 0.01 in the rectangular and fan

shaped-room respectively.

In the rectangular room, not only did the mean RASTI value increase by 0.03 at the

middle positions of the room, but also the mean value decreased 0.003 at the rear

positions of the room, because diffused sounds had a relatively lower intensity of sound

energy than specular sound reflections. In the fan-shaped room with sound diffusing

panels on the side walls, the mean RASTI value increased approximately 0.02 over all

positions. Therefore, sound diffusers were not an effective to improve the speech

intelligibility, but they are necessary to create smoothly decaying reverberant sound

energy for music quality (Doelle, 1972). In experiment 2, the same result was found in

fan-shaped room with the modified ceiling shape.

Figure 5-6 illustrates the RASTI variations in each room when a sound absorbent

material was added on the rear wall. The mean RASTI values changed by 0.09 and 0.07

in the fan-shaped (60°) and rectangular room respectively. The highest mean RASTI

value in each position of each room was found at the rear positions of the room. The fan-

shaped (60°) with the sound absorbent material on the rear of the room had the highest

mean RASTI value of 0.59, and the rectangular room had the mean RASTI value of 0.55

at the rear positions of the room. The mean RASTI values at the front and middle

positions of the rectangular room were 0.51 and 0.53 respectively.

For each different room (Figure 5-5), the same amount of the absorbent material

was added. The entire area of the rear wall surfaces in the rectangular room was covered

28

by the absorbent material, but only 1/4 part and 1/5 part of the rear wall were covered by

the absorbent material in the fan-shaped and the round-shaped room respectively.

However, they all had the same room volume and the same amount of absorption.

Therefore, if the different shapes of rooms have the same volume and seating capacity,

the room that has a relatively small surface area of the rear wall is the most favorable

design in terms of preventing later arriving sound energy from rear walls. In addition, in

the rectangular room, the sound absorbent material on the rear wall also can prevent

flutter echoes between the front and rear walls.

NF REC

Figure 5-4: Drawings of Composite acoustical treatment at fan shape room (60°) and recta

ngular room

REC NF

FS RND

Figure 5-5: Drawing of sound absorption material on rear walls. Different proportion of th

e absorption material to the rear wall depending on room shapes

29

ABS VS RASTI

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

Rectangular Fan shape (90')Fan shape(60') Round shape

Figure 5-6: Comparison of RASTI in different models with the absorption material at the

rear wall

Section Studies of Fan-Shaped Room (Experiment 2)

The first experiment showed that the fan shaped (60°) room model had the higher

mean RASTI values of 0.50 to 0.60, but the values were not high enough to provide

‘good’ speech intelligibility. According to Steeneken and Houtgast (1980), both RASTI

and STI values from 0 to 0.30 are ‘bad’, 0.30 to 0.45 are ‘poor’, 0.45 to 0.60 are ‘fair’,

0.60 to 0.75 are ‘good’, and 0.75 to 1 are ‘excellent’ subjective intelligibility scales.

Four different ceiling shapes were designed for the fan shaped (60°) room model.

Figure 5-7 illustrates the ceiling shape modifications of each room. All of these rooms

have identical room volumes of 5900m3 and seating capacity of 660 seats as the original

fan-shaped model (60°).

Figure 5-9 shows comparisons of RASTI values at 9 receiver positions in each

section model. Ceiling type A had the highest mean RASTI value of 0.5 at the rear

positions of the room. In this room, the mean RASTI values at the front and middle

positions increased by approximately 0.02 and 0.03.

30

Ray diagrams in figure 5-8, showed that reflected different sound distributions

occurred in different section models. Ray diagram of ceiling type A explains that the

sloped angle of this model is designed to distribute the sound energy to the middle and

the rear of the room. The ceiling height above the sound source was 2m higher than

ceiling type D.

Ceiling type B was also designed with a sloped ceiling angle at the front of the

room, but the angle of the ceiling was different than that used in ceiling type A. Ceiling

type B had a much smaller angle between the front wall and the ceiling than ceiling type

A. As shown figure 5-9, ceiling type B model had a mean RASTI value that was

approximately 0.02 lower than the original fan shaped room at the middle positions of the

room. A ray diagram of ceiling type B explains that not much reflected sound energy

arrived at the middle room compared to ceiling type A. The mean RASTI value in the

ceiling type B model was approximately 0.47 which was the same mean RASTI value as

the flat ceiling model.

In a ceiling type C, additional volume was added at the front room in order to

decrease the ceiling height over the rest of the room. However, this room had the lowest

mean RASTI value of approximately 0.43. The mean RASTI value was approximately

0.05 lower than the original fan shaped room at the rear positions of the room. This can

be explained by the lack of early sound reflections from surfaces around sound source

resulting in the decrease of speech intelligibility.

31

Ceiling type A Ceiling type B

Ceiling type C Ceiling type D

Ceiling type A + Additional treatments Ceiling type D + Additional treatments

Figure 5-7: Different section drawings of fan shaped room and additional treatments

(Canopy above the pulpit, sound diffusers on the side walls and absorbent materials on the rear wall) on surfaces.

32

Type A Type B

Type C Type D

Figure 5-8: Ray diagrams in the four different sections of fan shape room models. (With- out surface treatments)

Ceiling type D had a relatively lower ceiling height all over the room than ceiling

types A and B. This room had not only a short distance between the sound source and its

surrounded walls, but also a sloped ceiling above the sound source. Figure 5-9 shows that

the mean RASTI values at the front and the middle positions of the model with ceiling

type D were the same as the model with ceiling type A, but the ceiling type A had mean

RASTI values that were 0.1 higher than the model with ceiling type D at the rear

positions of the room.

It was concluded that the lower ceiling height above sound source may provide

higher speech intelligibility in most parts of room because it can help to provide strong

early sound energy reflections. However, if the ceiling height is not high enough to

distribute the sound reflections to the rear room, speech intelligibility at the rear of the

room will be poor relative to the other parts of the room.

33

As shown in figure 5-7, acoustic materials were added at the model with ceiling

type A. Figure 5-10 shows the comparisons of RASTI values with different surfaces

treatments. Sound diffusing panels were added on the lower part of the side walls. As

shown figure 5-10, the values were decreased in most parts of the room. Generally

diffused sounds have relatively low amplitude of sound energy level compared to

specular sound reflections, but they have the same amount of total sound energy (Siebein

& Cann, 1989). Since RASTI values are related to the early arriving sound energy,

diffused sounds may not produce strong early arriving sounds. That’s one of the reasons

why the mean RASTI values decreased with sound diffusion panels on the side walls.

RASTI vs Ceiling shape (fan shaped 60')

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

Flat Shape Ceiling Type A

Ceiling Type B Ceiling Type C

Ceiling Type D

Figure 5-9: Comparisons of RASTI values at different ceiling type of fan shaped room.

See the drawings of each modeling at figure 5-7

When a canopy was suspended over the pulpit area, the mean RASTI values

increased slightly from 0.471 to 0.477, but when the sound absorbent material was placed

on the rear of the room, the mean RASTI values increased to 0.497. The mean RASTI

values increased the most even though only a limited amount of sound absorbent material

34

was added to the rear wall. In this model, only 1/3 of the rear wall was covered with the

absorbent material. In addition, the mean RASTI values increased only 0.003 when sound

diffusing panels were placed on the side walls. This is the same result found in

experiment 1. However, as shown figure 5-9, the values were not high enough to indicate

‘good’ rating of the speech intelligibility quality (RASTI higher than 0.60) with all of

different treatments combined.

Acoustical treatment VS RASTI (NF-A model)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

without diff + low

Diff at low and high ABS at rear

Canopy total

Diff refers to sound diffusing panels on the sidewalls, ABS refers to the absorbent

material on the rear wall, canopy refers to the suspended canopy above the pulpit, and total refers to the composite design of previous treatments

Figure 5-10: Comparisons of RASTI values at various surface treatments at ceiling type

A for fan-shaped room

Section Studies of Rectangular Room (Experiment 3)

The rectangular room provided the second highest mean RASTI values of 0.44

among the three basic models in experiment 1 (Figure 5-1). Additional surface wall

treatments and room shape modifications were performed in the rectangular model.

Figure 5-11 illustrates various sections of the rectangular room with the acoustical

treatments studied.

35

Comparisons of RASTI values were performed in the rectangular room with a

concave shape ceiling (SCC), modified concave shape ceiling (SS), double concave

shape ceiling (DCC) and double convex ceiling (DCV) shape. Figure 5-12 shows that all

these shaped rooms had mean RASTI values over 0.60 which indicates a ‘good’ speech

intelligibility perception. In this study, the standard deviations of the RASTI values at

different positions were between 0.01 and 0.046. The single concave ceiling had mean

RASTI values of 0.63 and 0.64 at the front and rear positions of the room. However, in

the middle positions of the room, the mean value was 0.72. The standard deviation of the

RASTI values in this model was approximately 0.046.

In addition, both the double concave and triple concave shaped ceiling rooms

provided higher RASTI values than the double convex and triple convex shapes. It was

an unusual result because the concave shape may cause sound focusing problems and this

has been thought to decrease speech intelligibility in areas outside the focusing area.

To investigate these results, ray diagrams of each section model were compared. As

shown figure 5-13, the single concave ceiling room had the most uniform reflected sound

distributions especially middle seating areas. This is one of the reasons that caused the

highest RASTI values in the middle positions of the room. When compared with the

double concave ceiling (DCC) and the double convex ceiling (DCV) room, the ray

diagrams show that the ‘DCC’ had much more early reflected sound energy arriving in

the middle of the seating area than the ‘DCV’ room had at the same area.

The mean RASTI values in model type ‘DCC’ was 0.64, and in model type ‘DCV’

was 0.60. In the model type ‘DCC’, the highest mean RASTI values of approximately

0.65 were found at the middle positions of the room. The standard deviation of the

36

RASTI values in each model varied from 0.01 to 0.019 when the ‘SCC’ type model was

excluded. Model type ‘SCC’ had the highest standard deviation of the RASTI values of

approximately 0.045. In this model, the highest mean RASTI values 0.72 were found in

the middle positions. The mean RASTI values at the front and rear positions of the ‘SCC’

room were 0.62 and 0.63 respectively.

The trends of Early Energy Fraction (D50) in the single concave ceiling room are

shown at figure 5-14. The D50 values in octave bands were similar to the RASTI

variations in the rectangular room with single concave ceiling at the front of the room

(figure 5-12). It was found that the variation of RASTI values had very strong relation

with D50 indicating that early sound energy is important to improve speech intelligibility.

The sound pressure levels (SPL) in each position at each octave band were

investigated to figure out the relationship between possible sound focusing and the

speech intelligibility.

Figure 5-15, shows that the SPL at 1K had the most similar graph as the RASTI

variations in figure 5-12, but the difference between the highest mean SPL value at the

middle positions (48.7 dB) and the lowest mean SPL value at the rear positions (46.1 dB)

was only 2.1 dB. Egan (1988) stated that, a sound pressure level change in 3 dB is just

barely perceptible and 6 dB changes make the difference clearly noticeable.

If there was significant sound focusing in the middle of the room, the SPL values at

these positions should be higher than for the other locations in the room. Therefore, it

was concluded that there was not much major sound focusing at the middle positions, and

this was not the reason for the increasing RASTI values at these positions.

37

Single concave ceiling (SCC)

(Sound absorption on the rear wall)

Single concave + convex (SS)

(Sound absorption on the rear wall)

Double concave ceiling (DCC)

(Sound absorption on the rear wall)

Double convex ceiling (DCV)

(Sound absorption on the rear wall)

TCC (Sound absorption on the rear wall) TCC+RCV (Convex sound diffusers on the upper part the rear wall + sound absorption on the lower part of the rear wall)

TCV (Sound absorption on the rear wall) TCV+RCV (Convex sound diffusers on the upper part the rear wall and side walls + sound absorption on the lower part of the rear wall)

Figure 5-11: Different section drawings of rectangular room and additional treatments

38

Ceiling shape VS RASTI (Rectangular model)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

badFlat ceiling SCC SS

DCC DCV

RASTI comparisons convex VS concave

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

Single ConcaveS ( concave + convex )Two ConcaveTwo Convex

Figure 5-12: Comparison of RASTI values for rectangular room on various ceiling types

throughout 9 receiver positions at each room. In all these rooms sound absorption materials were added at the rear wall.

39

SCC DCC

DCV

Figure 5-13: Ray diagrams of different sections for the rectangular room.

The ceiling height above the pulpit area was changed to 4.8m, 5.8m, and 6.8m to

find an appropriate ceiling height to create maximize speech intelligibility. It was

expected that a decreased ceiling height may result in an increase in RASTI values

because of strong early reflections from the ceiling. However, as shown figure 5-16, the

ceiling height of 5.8m had the highest mean RASTI values of 0.65. In this model, the

mean RASTI values were approximately 0.66 at the front of the room while the other

rooms had similar RASTI values of approximately 0.61 at the front positions of the room.

The mean RASTI values in the model with ceiling heights of 4.8m and 6.8m model were

0.61 and 0.62 respectively. Therefore, it was concluded that the ceiling height above the

sound source must be decided pertinent to the room dimensions.

The ceiling height was also modified above the congregational seating area. Figure

5-17 compares the RASTI values depending on the room height to length ratio. Here, the

room length was set to 35m. As the ceiling height decreased, the mean RASTI values

increased at the front positions of the room. The ratios in the models with the ceiling

40

heights of 4.4m and 5.8m model had the highest mean RASTI values of approximately

0.65. The ratio 3.8 model with the ceiling height of 3.8m had the lowest mean RASTI

value 0.62. The highest mean RASTI value (0.67) in the middle positions of the room

was found in the model with a ceiling height of 4.4m. However, the highest mean value

at the front (0.67) and the rear positions (0.64) of the room was found in the model with a

ceiling height of 5.8m.

One of the preconditions of this experiment 3 was that the whole rear wall was

covered with sound absorbent material. For this reason, The RTmid decreased from 2.0s to

1.65s, because a majority of the sound energy was absorbed by both the congregation and

the rear wall. Siebein suggested that sound diffusive convex wall between the rear part of

the ceiling and the upper part of the rear wall is one of the ways to decrease the amount

of sound absorbent material required and to increase the reverberation time. Figure 5-18

shows that the mean RASTI values decreased from 0.66 to 0.64 or 0.62 when a convex

wall was placed between the ceiling and the rear wall instead of the whole absorption.

41

Receiver positions

Figure 5-14: D-50 (%) trend of the single concave ceiling room at each octave band

Receiver positions

Figure 5-15: Sound pressure levels (dB) trend of the single concave ceiling room

42

Ceiling height VS RASTI

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

6.8 m5.84.8

Figure 5-16: Comparison of RASTI for three different ceiling heights at above pulpit area

for the rectangular room

Ceiling height to room length ratio VS RASTI

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad3.18 3.5 3.8

4.4 5 5.8

Figure 5-17: Comparison of RASTI for different ceiling height (above congregation) to

length ratio for the rectangular room (Room length is always 35m)

43

Wall treatment at the rear wall VS RASTI

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Front Middle Rear

Receivers positions

RAS

TI

0

0.15

0.3

0.45

0.6

0.75

0.9

excel

good

fair

poor

bad

ABS on rear wall

Single convex diffusion on rear wall

Double convex diffusion at rear wall

Figure 5-18: Comparisons of RASTI values for rectangular room with different surface

area treatment; absorption and convex sound diffusive wall between rear part of ceiling and rear wall

CHAPTER 6 RESULTS AND ANALYSIS OF EXPERIMENT FOR CHURCH MUSIC

Music sound quality in a church is important to provide an optimum environment

for choir and congregational singing, and instrumental accompaniment that are an

important part of worship services.

Music sound quality can be evaluated by objective quantities or subjective

quantities. Objective quantities can be predicted by several acoustic parameters, and

subjective judgment can be evaluated by asking listeners for their preference of each

parameter in sound fields. However, Bradley (1995) concluded that ‘subjective

evaluation of new room acoustic measure’, several acoustic measurements were found to

have strong correlation with each subjective rating. In this study, an omni-directional

music sound source was located 3 m from the front wall at each model. Objective

evaluations were performed with several acoustic parameters which were obtained

through computational analysis. The average of at each parameter was compared with

three seats at each distance: front, middle and rear.

Loudness

The relative strength (G), dB was compared in different shaped models. Figure 6-1

shows that the front positions of each room had the highest G values and the rear

positions of each room had the lowest G values, because the sound intensity is attenuated

by geometric spreading due to increased distance from the source, the air molecules, and

absorbent materials. Generally as a sound receiver is located father away from the sound

sources, G values tend to decreased. However, in the rectangular room with single

44

45

concave ceiling (SCC), the mean G value at the front (8.4 dB) and middle of the room

(8.2 dB) only differed by approximately 0.2 dB. The G value at rear position decreased to

6.4 dB. According to the early directional echogram shown appendix D-1, this model had

very strong early reflections at the middle position of the room, and this is one of the

reasons why G values did not decrease much in middle of the room. In the narrow fan-

shaped room, the mean G value difference at the front (9.5 dB) and middle positions of

the room (7.5 dB) was approximately 2.5 dB.

The highest average G mid at each position was found in the fan-shaped model. In

the fan-shaped model, the front positions in the room had the highest average G mid values

of 10.3 dB. The middle and the rear positions of the fan-shaped room had G values of 8.7

and 8.0 dB respectively.

Figure 6-1 shows that the change of the ceiling shape from flat at the front of the

room to type A for the fan shaped increased G (AVG) at middle positions of the room

from 7 to 8.2 dB. Type SCC for the rectangular shape also increased G (AVG) at the

middle positions of the room from 6.3 to 8.1 dB. The highest standard deviation of the G

value (2.1 dB) was found in the rectangular room. The standard deviation decreased to

0.97 dB when the ceiling shape was changed to ‘SCC’ type model because the model

with a single concave ceiling could help to distribute the sound energy over all the room.

The standard deviation value also decreased from 1.3 to 0.75 when the ceiling shape

changed from the fan-shaped room with flat ceiling to the fan shaped with ceiling type A.

Normally, an increase of the G value 2 to 3 dB is a significant difference (Hidaka &

Beranek, 2000).

46

Where room size is between 1000 to 2000 seats, G mid 4 dB to 5.5 dB is desirable in

a concert hall (Siebein & Kinzey, 1998). In addition, it was noted that G mid is

proportional to reverberation time and inversely proportional to the room volume (V)

when the Sabine equation is valid (Hidaka & Beranek, 2000).

G values comparison

0

2

4

6

8

10

12

Front middle rearReceivers pos it ions

G (

dB)

REC RCC FSNF NF- A RND

Figure 6-1: Comparisons of G (dB) values in different shape room models and ceilings

with omni-directional sound source

Table 6-1: Average of relative strength G mid (dB) in different models (sound absorbent material on the rear wall)

FS REC RND

NF-A (Slope ceiling)

SCC (Single concave)

9 dB 7 dB 8.1 dB 8.9 dB 7.6 dB

Clarity

Clarity is the ability to hear musical notes or speech syllables as individual sounds.

The clarity index (C80) is the ratio between direct sound energy and early reflected sound

energy within the first 80 ms and the later or reverberant energy (Siebein et al., 1997).

Fig 6-2 shows that C80 values vary with room shape and receiver positions. The

rectangular room with single concave ceiling had the highest average C80 values vary

47

between 4.6 dB (Rear) to 7.2 dB (Middle). It was found that the impulse response in the

middle positions of the type ‘SCC’ had much higher sound pressure level at early

reflections than the flat ceiling model, ‘REC’. The early directional echogram in

Appendix B-1 shows that the model type ‘REC’ had maximum Lp of individual

reflections approximately 30 dB at first 50 ms, but the model type ‘SCC’, as shown

appendix D-1, had maximum Lp of individual 40 dB at first 50 ms. For this reason, C80

values appeared higher at the middle positions than the front and middle positions of the

‘SCC’ type model.

Figure 6-2: Distribution of C80 in different seats in the fan shaped and rectangular room

Table 6-2: The comparison of C80 (AVG) in different church room models measured in CATT Acoustic and preferred C80 at three different shape of model (Siebein & Kinzey, 1999)

NF REC RND NF-A SCC Concert hall Theater Classroom 2.9 1.5 -0.5 1.1 6 +1 to -4 (-6) 0 to +4 +5

Table 6-2 shows that C80 (AVG) in the fan-shaped and the round-shaped room

range within the preferred C80 for a theater room (0 to +4) and the rectangular room with

single concave ceiling had C80 (AVG) 6 dB that was similar to the acceptable value (+5)

for a classroom.

As shown in figure 6-3, the C80 (AVG) was increased approximately 3.1 dB when

the ceiling shape was changed from flat ceiling (REC) to the single concave ceiling shape

(SCC) at the front area of the rectangular room. It was found that the rectangular room

48

with a single concave ceiling had the highest clarity which was appropriate to speech

sound, but it is difficult to say that it also has an acceptable C80 for Catholic Church

music that traditionally is heard in a reverberant environment. In addition, the C80 values

variations were different in each model. The rectangular room had the lowest standard

deviation of C80 values 0.27 dB and the fan-shaped room had the highest standard

deviation of C80 values 1.1 dB. In generally, C80 values between 0 to -3 dB are preferred

for orchestral music (Hidaka & Beranek, 2000). Therefore, the round shaped room had

preferred C80 values at all three positions of the room.

C80 comparisons

- 5.0

- 3.0

- 1.0

1.0

3.0

5.0

7.0

9.0

Front Middle Rear

C80

(dB)

REC NF RND

NF- A SCC

Figure 6-3: Comparison of C80 (dB) values in different shape room models with omni-directional sound source

Reverberance

Sabine’s RT was measured in CATT Acoustic and directly used in this study. As

discussed in chapter 4, highly sound reflective materials were used for most room

surfaces and sound absorbent material was placed only on the rear wall in order to

provide a similar RT in different models. However, increasing the amount of absorbent

material resulted in decreasing of RT. Therefore, as suggested by Siebein, providing

49

highly diffusive surfaces at the rear ceiling and upper corner of the rear wall (Fig 5-11, 5-

18) may reduce the amount of absorbent material required and increase RT. In the

rectangular room with the single concave ceiling, The RT at 500(Hz) with absorbent

material on the rear wall was 1.76 s. It increased to 1.94 s when sound diffusing pannels

were added with fully occupied room condition. In Catholic Church room, the desired RT

is between 2.0 to 2.4s when it is fully occupied in 7000m3 room volume (Doelle, 1972).

Intimacy

Intimacy is the acoustic perception of sound such as it is heard in a small room. The

Initial Time Delay Gap (ITDG) is a factor that indicates the intimacy of a room and is

usually measured as the difference between the arrival time of the direct sound and the

first reflected sound (Egan, 1988). In this study, ITDG values were estimated using ray

diagram analysis. Figure 6-4 shows that the ITDG values vary at different locations of the

room: front, middle, and rear. All of the rooms have the similar variation of ITDG values.

The front positions had the highest ITDG values between 29 ms and 40 ms, the middle

positions had intermediate ITDG values between 17 ms and 30 ms, and the rear seats had

the lowest values between 11 ms to 23 ms. However, the rectangular room had the lowest

values inmost parts of the room, and the fan-shaped room had the highest ITDG values

all over the room. The Rectangular room had an ITDG value of approximately 17 ms at

the middle position of the room and 11 ms at the rear position of the room.

After changing the ceiling shape in the rectangular room with the flat ceiling to

model type ‘SCC’, the ITDG value was changed from 17 to 24 ms at the middle position

of the room. An ITDG between 20ms and 40 ms produces a perception of acoustic

intimacy. In concert halls an ITDG less than 20ms (at the center of the main floor) is

50

preferred (Hidaka & Beranek, 2000). Therefore, in this study, the middle and the rear

position of each room had ITDG values less then 20ms.

ITDG VS room shape

0

10

20

30

40

50

60

70

80

Front Middle RearReceivers positions

ITD

G (m

s)

REC RCC

FS RND

Figure 6-4: Comparisons of the estimate ITDG values at 3 receiver positions at each

model.

Spaciousness

Spaciousness is the perception of three dimensional sound fields. Spaciousness is

related to lateral sound reflections between 80 and 100ms from the sides and the rear of a

room (Siebein, 2001). The Interaural Cross Correlation (IACC) is a factor to predict

spaciousness. The IACC can be obtained through the ‘Auralization’ (Defined chapter 4)

procedure in CATT acoustics analysis. In this study, acoustic guitar sound from anechoic

sound sample in CATT was applied to the echogram file, which is created by the

prediction menu, and the IACC was calculated after the completion of the post processing

(CATT 8.0, Help).

The IACC values at 9 positions in the different shaped churches are compared in

figure 6-6. This figure estimated the average IACC values at the octave band frequency at

500 Hz, 1000 Hz, and 2000 Hz. A review of this figure shows that the IACC values were

similar in the rectangular and the fan-shaped room. The highest mean IACC value was

51

approximately 0.65 in the round-shaped room, and the lowest mean value was 0.36 in the

rectangular room. However, the IACC values varied at different positions in each model.

In the fan-shaped and round-shaped model, the IACC values increased as listeners

were located close to side walls, because as the listeners located farther away from the

center positions, the relative difference between the sounds that arrive at the left and right

ears also increased. In the fan-shaped room, the mean IACC value at three positions near

sidewalls was 0.36, but the mean IACC value at the three center positions of the room

was 0.5. Generally, IACC values less than 0.5 are desirable for most concert hall music

(Siebein & Kinzey, 1999).

Figure 6-5: Example of impulse response at both ears, L and R, at the center position of

Rectangular room with flat ceiling.

IACC (500,1K,2K) c omparis ons

0.000.100.200.300.400.500.600.700.800.901.00

Front Middle RearRec eivers pos it ions

IAC

C

REC NF RNDNF- A SCC

Figure 6-6: Comparisons of IACC values at the average of octave band frequency (500

Hz, 1000Hz & 2000Hz)

52

Warmth and Brilliance

Bass ratio and Treble ratio are factors that indicate the support of low frequency

sound and high frequency sound respectively. They are defined by the following

formulas.

Bass ratio : Treble ratio RT RTRT RT

125 500

500 1000

++

: RT RTRT RT

1000 2000

500 1000

++

In this study both factors are calculated by using the Sabine RT from CATT

Acoustic. However, all different shape churches had almost same Bass ratio and Treble

ratio which are 1.35 and 0.8 respectively, because they had identical RT mid

approximately1.9 s. A Bass ratio between 1.0 and 1.2 or even higher is desired and a

Treble ratio between 0.8 and 1.1 is appropriate for general concert hall music. In order to

provide a sense of acoustic warmth, the reverberation time at 125 Hz should be 10 to 50

% longer than the mid-frequency reverberation time. It also has been found that the

strength of the bass sounds must be higher than the mid-frequency sounds relatively

(Siebein & Kinzey, 1999).

Envelopment

Envelopment is the feeling of being enveloped by the sound field and is mostly

related to early sound energy arriving from the side walls within 80 ms.

In this study, the lateral energy fraction (LEF) was directly measured from CATT

acoustic. It was possible to predict the subjective feelings of envelopment in different

shaped rooms. As shown in table 6-4, LEF values varied with the geometry of the rooms.

At listener location close to the side walls, the LEF values increased, because, much more

sound energy arrived in this area from the near side walls than sound energy arrived at

the center of the room. The rectangular room had the highest mean LEF value of 0.25.

53

This model had the highest mean value at the middle (0.28) and front (0.25) positions of

the room among different shape models. The highest value at the rear positions (0.27)

was found in the fan shaped room. The mean LEF value at the rear positions of the

rectangular room was 0.23. The round-shaped room had the lowest value over all the

rooms. The average of LEF values at 9 positions was 0.08. As stated earlier, the round-

shaped room had the highest mean IACC values (Figure 6-3), and it had the lowest mean

LEF values. Therefore, it was concluded that the round-shaped room would be perceived

as the least spacious when it was compared to the rectangular and the fan-shaped room,

and it would be perceived as the least enveloping of the room. In this study, the LEF

values in all of the rooms were located within the preferred LEF values at the music room

regardless of their room shape.

As shown figure 6-7, when the front ceiling shape was modified in the rectangular

and the fan-shaped room, the average LEF values were decreased from 0.25 to 0.21 and

from 0.18 to 0.178 respectively. To increase lateral reflections, the side wall must be

designed to diffuse sound energy by irregular surfaces such as ornate decorations or

sculptures or to direct lateral reflections to seating locations (Hidaka & Beranek, 2000).

As shown table 6-4, all of models that were investigated in this study had

appropriate music qualities of reverberance, intimacy, spaciousness, envelopment,

warmth, and brilliance. However, C80 values were higher than preferred range for

musical sounds, but this provided appropriate range for speech clarity.

Table 6-3: The comparison of LEF of the average of 9 positions in different church room model measured in CATT Acoustic and preferred LF value (Siebein & Kinzey, 1999)

FS REC RND

NF-A (Slope ceiling)

SCC (Single concave)

PREFERRED

0.18 0.25 0.08 0.18 0.21 LEF< 0.40

54

LEF comparisons

0.000.100.200.300.400.500.600.700.800.901.00

Front Middle RearReceivers positions

LEF

REC NF RNDNF- A SCC

Figure 6-7: Comparison of LF values in different shape room models with omni-

directional sound

Table 6-4: Summary table of acoustical qualities in five models (AVG refers to average of each parameter, SD refers to standard deviation of each parameter, bold numbers refer to values within or close to preferred values in each parameter for music)

Acoustical Quality

Acoustical Measures REC SCC NF NF-A RND Preferred

Loudness G mid (dB) AVG 7 7.6 7.8 9. 8.1 4 ~ 5.5

SD 2.1 1 1.3 0.8 1.6

Clarity C80 (dB) AVG 1.5 6 2.9 1.1 -0.5 0 ~ (-3)

SD 0.27 0.9 0.8 1 0.71

Intimacy ITDG (ms) 17 24 30 . 21 ITDG ≤ 20

At middle SD 9.7 9.4 9.7 8.7 6.9 Spaciousness IACC AVG 0.37 0.44 0.42 . 0.65 IACC < 0.5

SD 0.13 0.16 0.19 . 0.15

Warmth Bass ratio 1.35 BR ≥ 1.0 or 1.2

Brilliance Treble ratio 0.8 0.8≤TR≤ 1.1 Envelopment LF AVG 0.25 0.21 0.18 0.18 0.08 LF <0.40

SD 0.37 0.05 0.08 0.08 0.04

Reverberant RT (sec) 1.79-2.2 2.0 ~2.4

CHAPTER 7 CONCLUSIONS

The intention of this study was to find the appropriate architectural features of

churches to maximize speech intelligibility without using sound reinforcement systems.

Three computer room models: rectangular, fan-shaped, and round-shaped were made in

order to investigate the relationship between RASTI values and room shape. It was found

that the speech intelligibility of church rooms was affected by the room shape and the

acoustical characteristics of the surface materials of the room.

The mean RASTI values in different shape models varied between 0.42 and 0.46

giving a ‘poor’ rating in the quality of speech intelligibility. The highest mean RASTI

value of 0.47 was found in the front positions of the fan-shaped model and the lowest

mean RASTI value of 0.40 was found in the rear positions of the round-shaped room.

The lowest standard deviation of the RASTI values of 0.01 was found in the rectangular

room. The fan-shaped and round-shaped room had standard deviation value of 0.04. This

research observed that the rectangular and the fan-shaped (60°) rooms provided the mean

RASTI values that were 0.02 to 0.03 higher than the round-shaped room when the rooms

had the same room section.

It was found that different acoustical treatments at room surfaces also affected

speech intelligibility. When sound absorbent material was added on the rear wall, the

highest mean RASTI values of 0.55 were found in the fan-shaped (60°) room where the

subjective quality of speech is judged ‘fair’. The canopy suspended over the pulpit area

increased the mean RASTI value to 0.48 (Rectangular) and 0.47 (Fan-shaped 60°).

55

56

However, the RASTI values in the rooms with flat ceilings never exceeded 0.60

regardless of their room shapes and acoustical treatments on their surfaces.

It was determined that ceiling shape modifications at the front of the room could be

the most efficient way of improving the speech intelligibility through the RASTI value

comparisons and the analysis of ray diagram at various room sections. As shown figure

7-1, the mean RASTI value increased approximately 0.20 to 0.33 when the front ceiling

shape was changed from the flat ceiling to the concave or convex shaped ceiling models.

Concave ceiling VS flat ceiling

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

1.00

Front Middle Rear

RAST

I

excel

good

fair

poor

badSCC TCC TCV

SCC refers to single concave ceiling, TCC refers to triple concave ceiling, and TCV refers to

triple convex ceiling ||||||||||||||||||||||refers to the total RASTI variations throughout experiment 1 and experiment 2

Figure 7-1: Comparisons of RASTI variations between flat ceiling and concave or convex

ceiling at the rectangular room

The mean RASTI values in the flat ceiling rooms were found between 0.44 and

0.57 which indicated ‘poor’ or ‘fair’ rating of speech intelligibility with several acoustical

treatments. However, once the front ceiling shape was changed to concave or convex in

57

the rectangular room, the mean RASTI values were increased to 0.62 to 0.66 which were

‘good’ ratings of speech intelligibility.

Ray diagram analysis confirmed that the single concave ceiling shape at the front of

the room was the most efficient way to provide early sound energy to the listeners, and

that was the reason why the highest mean RASTI value of 0.72 was found at the middle

positions of the single concave shaped room. Therefore, it was concluded that the early

sound energy reflections from the ceiling above the sound source made this ceiling shape

one of the most important design elements to improve the speech intelligibility of church

rooms.

Several acoustic parameters were compared to investigate the relationship between

the music qualities and room shapes. It was found that all of models had good acoustical

qualities of reverberance, spaciousness, envelopment, warmth, and brilliance that were

within the recommended range for music listening.

However, the rectangular model with a single concave ceiling had higher C80

values (AVG 6 dB) than suggested in literature for musical sound, but appropriate for

speech clarity. Optimum speech intelligibility and reverberant sound for the music could

be satisfied at the same time, because the rectangular room with the single concave

ceiling panel at the front of the room had both a ‘good’ rating of speech intelligibility --

the mean RASTI values were located between 0.63 and 0.72 at different positions -- and

the preferred reverberation time of 2.0s simultaneously.

In church acoustics, as stated earlier, the ceiling and walls around sound source at

the front of the room could be dedicated to improving the speech intelligibility because

when the ceiling shape was modified at the front of the room, the mean RASTI values

58

were increased the most. In addition, the middle and the rear part of the room could be

devoted to control the reverberation time and other music criteria for congregational

music and the singing of worship spaces, but this may need further studies to verify the

facts. The following studies are recommended for future research.

1. Investigate how the middle and the rear parts of the room features can be modified to provide an optimum acoustical environment for church music.

2. The appropriate size and the locations of the canopy to improve speech intelligibility in church acoustics.

3. The relationship between diffused sound energy and speech intelligibility.

4. Subjective evaluation of acoustic quality in the computer-generated models through out ‘Auralization’ analysis in acoustic programs such as CATT or ODEON

APPENDIX A DATA FOR SEVERAL ACOUSTIC PARAMETERS IN DIFFERENT SHAPE ROOM

MODELS

Table A-1. Computer modeling test input data and RT comparison

Octave band centre frequency (Hz)

Band(Hz) 125 250 500 1000 2000 4000 8000Sound Source dB 65 68 71 74 77 80 83Background Noise dB 44 37 31 27 24 22 21 Surface absorption coefficient

Plaster on brick (side walls) 0.01 0.02 0.02 0.03 0.04 0.05

Tile on floor 0.02 0.02 0.02 0.02 0.03 0.03

Surface features

congregation on wooden pew 0.57 0.61 0.75 0.86 0.91 0.86

Concrete block on rear wall 0.02 0.02 0.03 0.04 0.05 0.07

Hard plaster on ceiling 0.14 0.10 0.06 0.05 0.04 0.03

RT. (Sabine) s Rectangular 2.51 2.53 2.14 1.92 1.74 1.46 Fan shape (90') 2.42 2.42 2.03 1.81 1.64 1.4 Round Shape 2.44 2.46 2.1 1.87 1.71 1.47 Fan shape (60') 2.74 2.74 2.3 2.04 1.85 1.53

60

61

Table A-2. Comparison of RATI with absorption at rear wall Rectangular Fan shape (90') Fan shape (60') Round shape

Abs on entire 1/3 Abs on lower Abs on entire 1/4 Abs on lower Conditions rear wall level of rear wall rear wall level of rear wall Receivers locations from altar

RASTI

left 0.48 0.47 0.58 0.41 Front center 0.47 0.49 0.6 0.43 right 0.46 0.52 0.6 0.47 left 0.53 0.47 0.59 0.4 Middle center 0.53 0.46 0.6 0.46 right 0.52 0.45 0.57 0.44 left 0.55 0.47 0.67 0.43 Rear center 0.55 0.48 0.64 0.52 right 0.54 0.46 0.6 0.47 Mean value 0.51 0.47 0.61 0.45 Max - Min value difference (%) 0.09 0.07 0.1 0.12 Volume (m3) 5893 5880 5901 5908 Sabine RT(s) at1KHz 1.55 1.57 1.45 1.61

62

Table A-3. Comparisons of various acoustical parameters at different positions in three models

Features Location from altar

Distance from sound source (m)

EDT(s) D-50 (Hz)

SPL (dB)

RASTI STI

C-80(dB)

rear left 26.50 2.04 62.7 53.5 0.49 0.49 3.50 rear center 26.00 2.37 65.1 53.1 0.48 0.49 3.90 rear right 25.50 2.18 59.7 53.8 0.46 0.47 3.70 middle left 17.00 1.80 56.7 53.7 0.47 0.48 3.20 Rectangular middle center 16.50 1.93 62.2 53.4 0.47 0.48 4.10 shape middle right 16.20 1.90 62.7 54.3 0.47 0.48 3.00 front left 9.00 2.09 39.5 56.0 0.43 0.45 -0.70 front center 7.80 2.14 50.0 56.7 0.43 0.45 1.10 front right 7.00 2.11 56.7 58.2 0.42 0.44 1.90 rear left 17.40 3.11 43.2 53.5 0.44 0.44 -0.70 rear center 18.00 2.28 51.1 54.9 0.45 0.45 2.60 rear right 16.50 2.74 42.4 53.9 0.42 0.44 -1.10 middle left 13.00 2.37 37.7 54.9 0.40 0.41 0.80 Fan shape middle center 12.30 3.04 55.7 53.8 0.41 0.42 1.60 middle right 11.00 2.43 26.2 54.7 0.41 0.43 -1.20 front left 5.30 1.65 43.1 56.1 0.44 0.43 1.10 front center 6.20 2.58 54.2 55.5 0.47 0.46 1.80 front right 8.50 2.00 50.2 56.9 0.49 0.51 2.90 rear left 19.40 3.19 24.9 52.4 0.40 0.41 -2.60 rear center 18.00 3.37 35.4 52.8 0.49 0.47 -1.00 rear right 17.40 2.85 53.4 54.8 0.41 0.43 0.80 Round middle left 14.80 3.28 21.0 52.9 0.35 0.36 -2.90 shape middle center 13.20 2.91 31.0 54.5 0.44 0.43 -0.10 middle right 12.80 2.84 50.4 55.8 0.38 0.41 1.20 front left 9.30 2.84 17.2 55.9 0.36 0.38 -6.00 front center 7.90 2.86 23.7 56.8 0.39 0.39 -3.10 front right 7.10 2.57 54.4 60.0 0.44 0.45 2.10

63

Table A-4. Comparisons of RASTI values with several acoustical treatments for rectangular and fan shaped room

Rectangular Without treatment

ABS at rear

Diff at sides

Canopy above Composite

0.43 0.48 0.39 0.48 0.51 Front 0.43 0.47 0.42 0.48 0.53

0.42 0.46 0.42 0.46 0.53 0.44 0.53 0.45 0.48 0.57

Middle 0.44 0.53 0.44 0.49 0.56 0.44 0.52 0.43 0.48 0.54 0.47 0.55 0.49 0.49 0.58

Rear 0.45 0.55 0.46 0.48 0.57 0.43 0.54 0.45 0.47 0.56

Average 0.44 0.51 0.44 0.48 0.55

Fanshape60' Without treatment

ABS at rear

Diff at sides

Canopy above Composite

0.45 0.52 0.44 0.43 0.52 Front 0.47 0.55 0.48 0.47 0.57

0.47 0.56 0.48 0.48 0.58 0.42 0.53 0.46 0.44 0.57

Middle 0.44 0.54 0.44 0.46 0.56 0.41 0.53 0.42 0.43 0.55 0.53 0.62 0.52 0.55 0.63

Rear 0.5 0.59 0.48 0.52 0.56 0.47 0.55 0.49 0.5 0.57

Average 0.46 0.55 0.47 0.48 0.57

Table A-5. Longitudinal section studies of rectangular shape modeling Conditions: Absorption materials are used at rear walls. No sound diffusion panels are placed on the surfaces Ceiling Type A Ceiling Type B Ceiling Type C Ceiling Type D Ceiling Type E

Descriptions Flat Shape Single Concave SS ( concave +

convex ) DCC DCV

Receivers locations from altar RASTI

left 0.46 0.64 0.65 0.46 0.64Front center 0.46 0.63 0.64 0.46 0.63 right 0.45 0.61 0.61 0.45 0.61

left 0.48 0.73 0.67 0.48 0.73center 0.53 0.72 0.64 0.53 0.72

Middle right 0.51 0.71 0.61 0.51 0.71 left 0.5 0.65 0.63 0.5 0.65Rear center 0.5 0.63 0.62 0.5 0.63 right 0.51 0.63 0.62 0.51 0.63

Mean value 0.49 0.46 0.59 0.57 0.46

Maximum - Minimum 0.08 0.04 0.11 0.06 0.04

Volume (cubic meter) 6224 5801 5810 5780 5626

Sabine RT (s) at 1K Hz 1.82 1.55 1.55 1.53 1.5

64

65

Table A-6. Room height to length ratio ( Length is always 35m) Descriptions 3.18 3.5 3.8 4.4 5 5.8 Receivers locations from altar

RASTI

left 0.61 0.66 0.59 0.63 0.65 0.69 Front center 0.60 0.67 0.60 0.65 0.65 0.68 right 0.59 0.64 0.60 0.64 0.63 0.64 left 0.63 0.67 0.63 0.68 0.65 0.64 Middle center 0.62 0.64 0.63 0.67 0.63 0.64 right 0.62 0.65 0.66 0.67 0.65 0.65 left 0.59 0.61 0.61 0.64 0.64 0.64 Rear center 0.59 0.63 0.64 0.64 0.65 0.65 right 0.61 0.62 0.61 0.61 0.64 0.65

66

Table A-7. Comparisons of RAST of Rectangular shape model with various surface treatment ( Flat ceiling type)

Descriptions Without treatment

Hanging sound reflective panels on 1m set forth from side walls

Hanging sound reflective panels on ceiling and side walls

Horizontal absorption materials strip on lower level of rear walls

Horizontal absorption materials strip on middle and upper level of rear walls

left 0.43 0.45 0.45 0.46 0.46 Front center 0.43 0.44 0.44 0.44 0.46 right 0.43 0.42 0.43 0.44 0.45 left 0.43 0.44 0.47 0.46 0.48 Middle center 0.45 0.45 0.46 0.50 0.53 right 0.45 0.43 0.44 0.46 0.51 left 0.42 0.44 0.46 0.49 0.50 Rear center 0.42 0.46 0.47 0.48 0.50 right 0.42 0.46 0.48 0.48 0.51

67

Table A-8. Comparisons of RASTI values with different front and rear ceiling shape modifications

Descriptions

Three Convex +ABS at rear wall

Three Convex + Two Convex at upper rear

Three Concave + Two Convex at upper rear

Three Convex + Two Convex at upper rear + Convex side wall

Three Convex + Two Convex at upper rear

Receivers locations from altar RASTI

left 0.61 0.60 0.63 0.58 0.62 Front center 0.61 0.59 0.62 0.58 0.62 right 0.58 0.57 0.60 0.57 0.59 left 0.63 0.57 0.56 0.58 0.69 Middle center 0.59 0.56 0.52 0.54 0.69 right 0.59 0.57 0.54 0.54 0.68 left 0.63 0.58 0.59 0.54 0.63 Rear center 0.63 0.59 0.61 0.60 0.63 right 0.60 0.60 0.60 0.57 0.62 Mean value 0.61 0.58 0.57 0.57 0.64 Volume (cubic meter) 5700 5434 5591 5434 5434 Sabine RT (s) at 1K Hz 1.5 1.63 1.67 1.63 1.63

68

Table A-9. Comparison of speech intelligibility between concave shape and convex

shape of the front ceiling

Descriptions Single Concave

Triple Concave

Triple Convex

Single Concave + 1 Convex at rear

Single Concave + Double convex at rear

Receivers locations from altar

RASTI

left 0.64 0.60 0.62 0.62 0.59 Front center 0.63 0.61 0.62 0.62 0.60 right 0.61 0.59 0.58 0.59 0.56 left 0.73 0.70 0.66 0.69 0.68 Middle center 0.72 0.70 0.64 0.69 0.67 right 0.71 0.68 0.62 0.68 0.64 left 0.65 0.63 0.62 0.63 0.61 Rear center 0.63 0.64 0.62 0.63 0.61 right 0.63 0.61 0.62 0.62 0.61 Mean value 0.66 0.64 0.57 0.64 0.62 Maximum - Minimum value difference

0.12 0.11 0.08 0.10 0.12

Volume (cubic meter)

5801

5877

5760

5651

5683

Sabine RT (s) at 1K Hz

1.55

1.57

1.54

1.66

1.72

APPENDIX B DATA FROM CATT ACOUSTICS-BASIC MODELS

Figure B-1. Detail data of position 06 without acoustical treatment

70

71

Figure B-2. Detail data of position 06 with absorption material on the rear wall

Figure B-3. Detail data of position 06 with canopy above the pulpit

72

Figure B-4. Detail data of position 06 with sound diffuser on the side walls

Figure B-5. Detail data of position 06 with composite acoustical treatment

73

Figure B-6. Detail data of position 02 without acoustical treatment

74

Figure B-7. Detail data of position 06 without acoustical treatment

75

Figure B-8. Detail data of position 06 with absorption material on the rear wall

Figure B-9. Detail data of position 06 with canopy above the pulpit

76

Figure B-10. Detail data of position 06 with sound diffusers on the side walls

Figure B-11. Detail data of position 06 with composite acoustical treatments on surfaces

77

Figure B-12. Detail data of position 09 without treatment

APPENDIX C DATA FROM CATT ACOUSTICS OF FAN-SHAPED ROOM

Figure C-1. Drawing of modified fan shaped room – ceiling type A

Figure C-2. Drawing of modified fan shaped room – ceiling type B

79

80

Figure C-3. Drawing of modified fan shaped room – ceiling type C

Figure C-4. Drawing of modified fan shaped room – ceiling type D

81

Figure C-5. Drawing of modified fan shaped room – ceiling type A + composite

acoustical treatments

Figure C-6. Drawing of modified fan shaped room – ceiling type D + composite

acoustical treatments

APPENDIX D DATA FROM CATT ACOUSTICS OF RECTANGULAR SHAPED ROOM WITH

MODIFIED FRONT CEILING SHAPE

Figure D-1. Drawing of modified rectangular room – ceiling type SCC

Figure D-2. Detail data of position 06 with sound absorbent material on the rear wall

83

84

Figure D-3. Drawing of modified rectangular room – ceiling type SS

Figure D-4. Detail data of position 06 with sound absorbent material on the rear wall

85

Figure D-5. Drawing of modified rectangular room – ceiling type DCC

Figure D-6. Detail data of position 06 with sound absorbent material on the rear wall

86

Figure D-7. Drawing of modified rectangular room – ceiling type DCV

Figure D-8. Detail data of position 06 with sound absorbent material on the rear wall

87

Figure D-9. Drawing of modified rectangular room – ceiling type TCC + RCV

Figure D-10. Detail data of position 06 with sound absorbent material on the rear wall

88

Figure D-11. Drawing of modified rectangular room – ceiling type TCV + RCV

Figure D-12. Detail data of position 06 with sound absorbent material on the rear wall

LIST OF REFERENCES

Acoustic Sciences Corporation. (2002). Common church acoustic problems & how to identify them [On line]. Available from URL: http://www.church-acoustics.com/problems.htm, (December, 2002)

Barron, M. (1988). Subjective study of British symphony concert halls. Acoustica, Vol. 66, No. 4, 1-14.

Barron, M. & Lee, L-J. (1988, April). Energy relations in concert auditoriums. I. J. Acoust. Soc.Am., Vol. 84, No.2, 618-628.

Berry, R. & Kinzey, B. Y.(1954, December) Planning for sound in church worship. Architectural Forum, 164-166.

Bradely, J. (1998). Relationships among measures of speech intelligibility in rooms: J. Audio Eng. Soc., Vol. 46, No.5,

Bradely, J. (1986, July). Auditorium acoustics measures from pistol shots. J. Acoust. Soc.Am., Vol. 80, No.1, 199-203.

Bradely, J. (1983, June). Experience with new auditorium acoustic measurements. J. Acoust. Soc.Am., Vol. 73, No.6, 2051-2058.

Bradely, J. & Soulodre, J. (1995, July). Subjective evaluation of new room acoustic measures. J. Acoust. Soc.Am., Vol. 98, No.1

Buglio, D. (1992). Why is church sound so confusing? [On line], Available from URL: http://www.jdbsound.com/art/art509.htm, (November, 2002) JdB Sound, Acoustic. Lab

D’Antonio, P. (1988). Acoustical control of worship spaces. Technologies for Worship Magazine [On line] Available from URL: http://www.jdbsound.com/art/art521.html, (December, 2002). JdB Sound, Acoustic. Lab.

David L. (1996, January). The distributed column sound system at Holy Cross Cathedral, Boston, the reconciliation of speech and music. J. Acoust. Soc.Am., Vol. 99, No.1, 417-425.

Doelle, L. (1972). Environmental Acoustics. McGraw-Hill, Inc., New York.

89

90

Egan, D. (1988). Architectural Acoustics. McGraw-Hill, Inc., New York.

Gilbert,A. & Soulodre, Bradley, S. (1995) Subjective evaluation of new room acoustic measures. J. Acoust. Soc. Am. Vol.98, No.1

Hawkes. (1971). Experiences in concert auditoria. Acoustica, Vol. 24, 236-250.

Hidaka, T. & Beranek L. L.(2000, January). Objective and subjective evaluations of twenty-three opera houses in Europe, Japan, and the Americas. J. Acoust. Soc. Am., Vol. 107, No.1.

Hidaka, T., Beranek L., Mauda, S., Nishihara, N., & Okano, T. (2000, January). Acoustical design of the Tokyo Opera City (TOC) concert hall, Japan, J. Acoust. Soc. Am., Vol. 107, No.1.

Ingram, D., Barnes, E., Hampton, C., Klepper, D., & Noehren, R. Acoustics in Worship Spaces: American Guild of Organists (pamphlet available from AGO, 815, 2nd Ave, New York, NY 10017)

Jarzabkowski, M. (2000). The role of acoustics in preliminary sanctuary design: Church Business, Build-in Sound, [On line] Available from URL: http://home1.gte.net/mjarzo, (December 2002)

Knudesn, V. & Harris, C. (1950). Acoustical Designing. John Willey & Sons, Inc., New York.

Kuttruff, H. (1991). Room Acoustics. Elsevier Sciences Publishers Ltd. New York.

Latham, H. (1979). The signal-to-noise ratio for speech intelligibility-- An auditorium acoustics design index. Applied Science Publishers Ltd. England

Lubman, D.(1983, November). Acoustics of Worship Spaces. American Institute of Physics for the Acoustical Society of America

Nagata (2001). Church acoustics. Nagata Acoustics News, Vol. 01-9, No.165.

Okano, T., Beranek, L. & Hidaka, T. (1998, July). Relations among interaural cross-correlation coefficient, lateral fraction, and apparent source width in concert halls. J. Acoust. Soc.Am., Vol. 104, No.1, 255-256.

Okano, T., Beranek, L. & Hidaka, T. (1995, August). Interaural cross-correlation, lateral fraction, and low-and high-frequency sound levels as measures of acoustical quality in concert halls. J. Acoust. Soc.Am. Vol. 98, No.2, 988-1007.

Reich, R. & Bradley, J. (1998). Optimizing classroom acoustics using computer model studies. Canadian Acoustics, 26(4) 15-21.

Riedel, R. (1986). Acoustics in the Worship Spaces. Church Music Pamphlet Series.

91

Rienstra, R. (1955, December). Church design for music, Architectural Record. 193-194

Riedel, R. (1983 May, 1984 May, 1986 January, 1987 May, 1988 April, 1990 April, and 1991 July). Acoustics in the worship space, I, II, III, IV, VI, VII, respectively. Diapason Magazine, 380 E. Northwest Highway, Suite 200, Des Plaines, IL.

Schroeder, M., Gottlob, D. & Siebrasse, K. (1974, October). Comparative study of European concert halls: correlation of subjective preference with geometric and acoustic parameters. J. Acoust. Soc.Am., Vol. 56, No.4, 119-125.

Siebein, W. (2002). Acoustical Design Issues in Contemporary Worship Spaces. Siebein Associates, Inc., Gainesville, FL.

Siebein, W. & Cann, G. (1989). Acoustical Modeling Workshop, A workshop presented at the ACSA 1989 Summer Institute on Energy and Environmental Controls

Siebein, W. & Crandell C., Gold A. (1997). Principles of Classroom Acoustics. Reverberation. Educational Audiology (Monograph 5).

Siebein, W. & Kinzey, Y. (1999). Architectual Acoustics. John Wiley & Sons, Inc. New York. 233-304.

Steeneken, H. & Houtgast, T. (1980, January). A physical method for measuring speech-transmission quality. J. Acoust. Soc. Am., Vol. 67, No.1.

Yoichi A. & Dennis N. (1995). Music and Concert Hall Acoustics. Academic Press. Harcourt Brace & Company, Publishers. San Diego.

BIOGRAPHICAL SKETCH

Sangjun Lee was born at Daegu, South Korea, in the year of 1975. In his early

twenties, he spent 26 months in military service (January 1995 to March 1997) in Korea.

After he completed the service, he traveled west Australia for 7 months and then back to

school. He received a Bachelor of Architectural Engineering degree (February 2001)

from the Kyungpook National University in South Korea where he emphasized the study

of architectural design.

He further pursued graduate study in architectural acoustics and attended the

graduate program at the University of Florida, entering in August 2001. During the two

years of his graduate study, he focused on the relationship between architectural design

and acoustics for developing fundamental design issues on architectural acoustics. In the

fall of 2002, he started his professional career as an intern acoustic consultant at the

Siebein Associates, Inc., Gainesville, Florida.

Upon graduation from the University of Florida, he intends to continue his research

on architectural acoustics for his Ph.D degree at the University of Nebraska from the fall

of 2003

92